Validity of a questionnaire to assess fruit and vegetable intake in adults

Transcription

1 (2006) 60, & 2006 Nature Publishing Group All rights reserved /06 $ ORIGINAL ARTICLE Validity of a questionnaire to assess fruit and vegetable intake in adults AG Kristjansdottir 1, LF Andersen 2, J Haraldsdottir 3, MDV de Almeida 4 and I Thorsdottir 1 1 Unit for Nutrition Research, Landspitali-University Hospital & Department of Food Science, University of Iceland, Reykjavik, Iceland; 2 Department of Nutrition, University of Oslo, Oslo, Norway; 3 Department of Human Nutrition, Royal Veterinary and Agricultural University, Copenhagen, Denmark and 4 Faculty of Nutrition and Food Sciences, University of Porto, Porto, Portugal Objective: To validate a method for assessing fruit and vegetable intake among adults in a population of low intake. The method assesses mean intake and ranks individuals by their usual intake. Design: A precoded fruit and vegetable Questionnaire included a 24-h recall and a food frequency. The participants filled in the Questionnaire, a week later they started a 7-day food record, 1-day weighed record and 6 days using household measures. Subjects: Following advertisements 40 participants were recruited, 36 returned food records(mean age ¼ 37 years). Results: No difference was observed between the average intake yielded by the 24-h recall and that from the 1-day weighed food record of fruits or vegetables. Correlation coefficients between results from the food frequency questionnaire and the 7-day food records were 0.45 (P ¼ 0.007) for vegetables, 0.63 (Po0.001) for fruits and 0.73 (Po0.001) for fruits and vegetables combined. Cross-classification into quartiles showed that the proportion of participants in the same or the adjacent quartile of the intake distribution were 94% for fruit intake and 80% for vegetable intake. Registered intake of fruits and vegetables was higher the first 4 days of the record than the last 3 days (P ¼ 0.002). The 4-day food record correlated with the food frequency questionnaire in similar manner as the 7-day record. Conclusion: The present study indicates that the precoded 24-h recall may be valuable tool for measuring average intake of fruits and vegetables among adults in a population of low intake. Moreover, the food frequency questionnaire was valid for ranking individuals according to their usual intake. A 4-day food record might be sufficient when validating food frequency questionnaires for fruits and vegetables. (2006) 60, doi: /sj.ejcn ; published online 23 November 2005 Keywords: fruit; Iceland; validity; vegetable; questionnaire Introduction A diet high in its variety of fruits and vegetables is associated with decreased risk of cardiovascular diseases and several types of cancer (Steinmetz and Potter, 1996; Ness and Correspondence: Professor I Thorsdottir, Unit for Nutrition Research, Landspitali-University Hospital & Department of Food Science, University of Iceland, IS-101 Reykjavik, Iceland. ingathor@landspitali.is Guarantor: I Thorsdottir. Contributors: AGK worked on the statistical analysis and wrote the first draft of the manuscript and made the greatest contribution to this paper. LFA, JH and MDVdA participated in designing the study and project planning. LFA and JH also participated in the data analysis. IT was the local project leader and participated in all parts of the work. All investigators contributed to the writing of the final paper. Received 26 April 2005; revised 16 August 2005; accepted 14 September 2005; published online 23 November 2005 Powles, 1997; Joshipura et al., 1999, 2001; Willett, 2000; Key et al., 2002; Hu, 2003). The health benefits of fruits and vegetables seen in epidemiology studies are the main reasons for the recommended intake of at least 400 g of fruit and vegetable per day, potatoes not included (WHO, 2003). Large population groups, particularly in Northern Europe, however, eat far less than the recommended amount of vegetables and fruits. The consumption is especially low in Iceland (The Norbagreen 2002 study, 2003; Steingrimsdottir et al., 2003; FAO statistic database, 2004). Although the consumption of vegetables, fruits and fruit juice has increased slightly between 1990 and 2002, which was significant for vegetables (Steingrimsdottir et al., 1991, 2003), it is far from the recommended intake and the Icelandic nutritional goal for fruits and vegetables of 500 g per day of vegetables, fruit and fruit juice. The mean consumption according to the

2 last national survey 2002 was 230 g of vegetables, fruits and fruit juice per day (Steingrimsdottir et al., 2003). A valid method is needed to evaluate the consumption, and especially changes in consumption level, where it is especially low. This paper describes a validation of an instrument aimed to be used in the Pro Children Project, where Iceland was one of nine participating European countries. The Pro Children Project was designed to provide information on actual consumption levels of vegetables and fruits in European school children and their parents, and to understand the determinants of consumption patterns among the children (Klepp et al., 2005). Moreover, the Pro Children Project aims to develop and test strategies, applicable across Europe, for promoting consumption of vegetables and fresh fruits among school children and their parents. Adults intake, that is, parents intake, has been found to be one of the key correlates of children s fruit and vegetable intake (De Bourdeaudhuij et al., 2005). An instrument to assess fruit and vegetable intake across Europe needed to be developed. Such an instrument should be brief, self-administrable, and give comparable results from countries at different consumption level. Moreover, the instrument should give valid information on the ranking of individuals according to their usual intake and on the mean intake of the group. Such an instrument is currently not available, that is, able to assess both group mean as well as ranking individuals from different countries, especially no instrument has been developed for low consumption population (Kim and Holowaty, 2003). The instrument developed in the study now presented included a food frequency part and a precoded 24-h recall. The food frequency part was used for ranking subjects according to their usual intake and the 24-h part was used for measuring group mean intake. Multiple-day food records have commonly been used for validation of food frequency questionnaires. The records should be kept for a sufficient number of days to represent the average intake (Cade et al., 2002). Cost and respondent burden has to be taken into consideration and Stram et al. (1995) have found that in most settings the optimal study design will rarely require more than 4- or 5- day diet records per subject. In the present study, the method s estimated mean intake was validated with 1-day weighed record and the method s estimation of usual intake was validated with 7-day food record. A shorter food recording period, that is 4 days, for validation of a food frequency questionnaire for fruit and vegetable intake was also evaluated. The objective of the present study was to study the validity of a Questionnaire for assessing fruit and vegetable intake among adults in a population of low intake. The method was aimed at assessing mean intake of a group and to rank individuals according to their usual intake. Subjects and method Subjects and design Adults were recruited through advertisements at three small work places and one small sport club in Reykjavik. A total of 40 adults agreed to participate, 13 men and 27 women. A total of 36 returned the records, 11 men and 25 women (participation rate 90%, mean age 37 years, range years). Participants received oral and written information about the study. All participants data were treated anonymously. The Icelandic Data Protection Authority approved the study protocol. The Questionnaire was validated against food records. The Questionnaire was answered a week before recording of the diet. The 24-h recall part of the Questionnaire was validated against a 1-day weighed record, and the food frequency part was validated against a 7-day food record, the first day with weighed food and the following 6 days with household measures. Prior to the validation a pretest had been sent home to parents of 11-year-old school children in Reykjavik (n ¼ 91) to test clarity and ease of completion. A total of 70 filled out Questionnaires were returned (participation rate 77%). None of these participants were included in the validation study. Structure of the questionnaire A Questionnaire was developed from validated tools used to measure the effects of fruit and vegetable subscription in schools in Denmark and Norway (Eriksen et al., 2003; Andersen et al., 2004). The optically scannable Questionnaire was composed of a 24-h recall part and a food frequency part. The 24-h recall part of the Questionnaire was included to give information about intake of the group and the amount and types of fruits and vegetables, whereas the food frequency part made it possible to rank individuals according to levels of usual intake. Only fresh fruits were defined as fruits and 100% pure juice as juice. From a nutritional point of view, fresh fruit are often low energy-dense foods relatively rich in vitamins, minerals and other bioactive compounds as well as being good source of fibers (World Cancer Research Fund, 1997). When canned, etc. sugar are usually added, and when fruit is dried it is not low energy-dense food anymore. This is the reason for fresh fruit being the main object of most countries recommendations about fruits and vegetables and the main object of the planned intervention for school children in the Pro Children project. Additionally, the usage of fruits in other than fresh format is negligible in Iceland, especially among the young population where it is hardly found (Steingrimsdottir et al., 2003). Potatoes were not included in the calculations of vegetable consumption. There was a question about the intake of potatoes in the food frequency part to prevent participants from marking potatoes as a cooked vegetable. In the 24-h recall part, the day was divided into three periods, that is, morning/midmorning, lunch/afternoon and 409

3 410 dinner/later in the evening. In each of the three periods of the day, participants were asked, using an open question, what they had eaten during that period the day before. This was done to prompt the participant s memory to that specific period of the day. Then there were specific precoded questions on juice, fruits and vegetables. The question on juice was: what kind of juice did you drink and how many glasses. The question on fruits was: what kind of fruit did you eat and how many s. Figure 1 shows an example of the question on fruits in the 24-h recall. Vegetables were further categorized into raw vegetables, salad and cooked vegetables. The question on raw vegetables was: what kind of raw vegetables did you eat and how many s or slices. The questions on salad and cooked vegetables were similarly: what kind and how many portions. Amounts were indicated in terms of the number of s, slices or portions eaten (Table 1). The portion sizes are supported by a recent paper from WHO (Agudo, 2005). The questions on salad and cooked vegetables were not included in the first period of the day as they are usually not consumed in this period of the day in the countries involved in the survey. Vegetable in composite dishes is thought to be of small importance in the total European diet, except the vegetable soup in Southern Europe. This may differ between individuals, but the amount H3. Did you eat fruit yesterday morning/midmorning? Yes No If yes, What kind of fruit? How much? Write 1 if you ate one apple, ½ if you ate a half Apple Banana Orange of vegetables in composite dishes is too hard to evaluate to be included in a simple questionnaire. To keep it as simple as possible, the 24-h recall part did not include specific questions on vegetables as part of composite dishes, except for vegetable soup. The food frequency part of the Questionnaire included six questions about the usual consumption of fruit, salad, raw vegetables, (potatoes), cooked vegetables and orange juice. The precoded response categories ranged from never to daily, more than twice a day. Figure 2 shows an example of a question on fruits in the food frequency part. Standards used for estimating the amounts in grams per day are shown in Table 2, supported by recent paper from WHO (Agudo, 2005). Food records Participants were asked to record food intake for 7 days, one day of weighed record and the following 6 days with household measures to reduce the burden on respondents. Table 1 Definition of portion sizes for the 24-h recall part of the Questionnaire used in the cross-sectional study Food item Portion size (g/portion) Fruit Apple, banana, orange, pear (s) 100 Tangerine () 50 Melon (slice) 50 Fruit salad (portion) 100 Other fruits (s/portion) 100 Fruit juice 200 Raw vegetables Tomato () 50 Cucumber (slice) 10 Carrot () 65 Other raw vegetables (portion) 50 Salad 40 Cooked vegetables 60 Vegetable soup 80 a a 80 g per 250 g soup. Tangerine Pear F1. How often do you usually eat fresh fruit? Melon Fruit salad Other fruit (write which) slice portion or portion Figure 1 An example of a specific precoded question on fruits in the 24-h recall part. Never Less than one day per week One day per week 2-4 days a week 5-6 days a week Every day, once a day Every day, twice a day Every day, more than twice a day Figure 2 An example of a specific precoded question on fruits in the food frequency part.

4 Table 2 Definition of portion sizes for the food frequency part of the Questionnaire used in the cross-sectional study (one portion at a time) Food item For the 1-day weighed food record, the participants were provided with a scale (PHILIPS HR 2385, Austria), and given written and oral instructions on how to use it, and how to record the diet. The first recording day was defined for each individual by the researchers in order to achieve an even distribution of weighed food records between Sunday, Mondays, Tuesdays, Wednesdays and Thursdays (the weekdays covered by the 24-h recalls). Three participants did not complete the 7-day record, one returned a 5-day record, and two returned 6-day records. For these three participants, a daily mean consumption was calculated only from the recorded days. Statistic The data were analysed using the computer program SPSS version 11 (SPSS Inc., 1999). For data analyses, the data from the questionnaire and the estimated food records were recoded into portions. The normality was tested with Kolmogorov Smirnov test. The distribution of the data was normal; hence, parametric statistical methods were chosen, except for juice, where non-parametric tests were used. The sample mean and standard error of the mean are presented. The differences between the methods were tested using the Student s t-test, Mann Whitney test or Wilcoxon s signed ranks test. The suitability of the food frequency for ranking participants was assessed by Pearson or Spearman correlation analysis and by classifying participants into quartiles. Results Portion size (g/portion) Fruit 100 Orange juice 200 Raw vegetables 50 Salad 40 Potato 70 Cooked vegetables 60 Table 3 shows mean fruit, vegetable and juice consumption (in g/day), estimated from the 24-h recall part of the Questionnaire and from the 1-day weighed food record. The mean intake estimated with the two methods did not differ significantly. The recorded intake of fruits, vegetables and juice diminished gradually from the first to the last day of the recording period. There was a significant difference of fruits and vegetables combined, between the first 4 days, and from the fifth to the last day of the recording period (P ¼ 0.008), as well as for fruit, vegetables and juice combined (P ¼ 0.002). Results from both the whole record period (approx 7-day food record) and from the first 4 days of the recording period are presented. Table 3 Intake of fruits, vegetables and juice in g/day, based on the 24- h recall part of the Questionnaire and the weighed food record (WFR), the first day of the 4-day food record (n ¼ 31 a ), mean (s.e.) Table 4 shows the mean consumption of fruits, vegetables and juice in g/day, estimated from the food frequency part of the Questionnaire and the 7-day food record as well as the 4- day food record. There was no significant difference between mean intake according to the food frequency part and the food record (7- and 4-day) for both fruit and vegetable intake; however, juice intake estimated from the food frequency part was significantly higher than from the food record (7- and 4-day). All 100% pure juice consumed by the participants during the recording period was 100% pure orange juice. There was no significant difference between mean intake in 7- and 4-day records for fruits, vegetables and juice separately, but when these food items were combined the 4-day record gave higher intake values. Correlation coefficients between results from the food frequency part of the Questionnaire and the 7-day food record were 0.45 for vegetables, 0.63 for fruits and 0.75 for fruit, vegetables and juice combined. Similar coefficients were found between the food frequency questionnaire and the 4-day food record (Table 5). Correlation between 7- and 4-day records was high. Cross-classification of participants by quartiles of calculated intake from the 7-day food record and the food frequency part of the Questionnaire showed that, overall, 94% of the adults were placed in the correct part of the consumption distribution for fruits (same plus adjacent quartile); for total vegetables this figure was 80% (Table 6). The food frequency part of the Questionnaire grossly misclassified 6% according to fruit intake and 20% according to vegetable intake. Similar results were found between the 4-day food record and the food frequency questionnaire. Compared with 7-day record, 4-day record had 97% or more of the participants correctly classified or in the adjacent quartile. Discussion 24-h recall WFR Difference Fruits 125 (19) 157 (24) Vegetables 99 (16) 105 (18) Juice b 87 (23) 90 (32) Fruits and vegetables 224 (21) 262 (31) FVJ c 311 (32) 352 (39) a 31 weighed the first day. b Student s t-test, except for juice, where Mann Whitney test was used (nonparametric test). c Intake of fruits, vegetables and fruit juice. The present study was conducted to validate a precoded Questionnaire in a population with a low intake of fruits and vegetables. 411

5 412 Table 4 Intake of fruits and vegetables in g/day, based on the food frequency part of the Questionnaire (FFQ) and the 7- and 4-day food record (FR), (n ¼ 36 a ) mean (s.e.). The differences between the three methods are shown FFQ FR (7d) Difference between FR (7d) and FFQ FR (4d) Difference between FR (4d) and FFQ Difference between FR (7d) and FR (4d) Fruits 103 (15) 109 (15) (15) Vegetables 82 (9) 83 (9) (11) Juice 105 (21) 64 (16) 41* (18) 31* Fruits and vegetables 187 (19) 186 (20) (21) * FVJ c 295 (31) 252 (25) 43* (28) * a n ¼ 36, except for vegetables as one participant did not fill in the food frequency part for vegetables. b Student s t-test, except for juice, where Wilcoxon s signed ranked test was used (nonparametric test). c Intake of fruits, vegetables and fruit juice. * Significant difference. Table 5 Correlation coefficients (Pearson or Spearman) between calculated intake from the food frequency part of the Questionnaire and the 7- and 4- day food record, and between 7- and 4-day records (n ¼ 36 a ) Correlation between FR (7d) and FFQ Correlation between FR (4d) and FFQ Correlation between FR (7d) and FR (4d) Fruits 0.63 o o o0.001 Vegetables o0.001 Juice 0.71 o o o0.001 Fruits and vegetables 0.73 o o0.001 FVJ c 0.75 o o o0.001 a n ¼ 36, except for vegetables as one participant did not fill in the food frequency part for vegetables. b Parametric test (Pearson correlation) except for juice where nonparametric test was used (Spearman correlation). c Intake of fruits, vegetables and fruit juice. The 24-h recall part of the Questionnaire gave valid intake data on the group mean intake level, compared with the weighed food record. There was no significant difference between intake estimated from the 24-h recall part of the Questionnaire compared with the 1-day weighed food record, that is, the first day of the record period. The mean fruit and vegetable intake according to the 24-h recall part of the Questionnaire was also similar to the 7- and 4-day food record and to the food frequency part of the Questionnaire. There are only a few validation studies on the 24-h recall method, focusing on fruit and vegetable intake. A 24-h recall instrument asking about fruit and vegetable intake similar to the one validated in the present study was validated against precoded food diaries, among 12-year old Norwegian children (Andersen et al., 2004). The 24-h recall overestimated the intake of fruit and fruit juice compared to the reference method, while there was no significant difference observed for vegetable intake. In an American study among adults, a telephone-administered focused-recall for fruit and vegetable intake was validated against a traditional 24-h recall (Neuhouser et al., 2001). The focusedrecall showed lower vegetable intake by 4.5% and higher fruit intake by 12.5%. The food frequency part of the Questionnaire was found valid and suitable for ranking individuals according to fruit and vegetable intake, when compared to dietary records. Diet records have been widely used and accepted as a preferred reference instrument in this type of study, performed to validate simpler methods like 24-h recall and short food frequency questionnaires, as recording involves errors that are largely independent compared to the other methods. A diet record does not rely on memory as do 24-h recall and food frequency questionnaires, nor does it rely on individuals perception of portion sizes, when it is a weighed record (Willett, 1990; Cade et al., 2002). Recording, however, requires the participants patience, time and work. Prolonged recording has been shown to lead to diminished recorded intake, and the amount of fruit and vegetables recorded is less after 4 days of recording (Andersen et al., 2002). In the present study, a significantly lower registered intake was observed on the last 3 days, compared with the first 4 days of the recording period. In the very beginning of the recording period, participants possibly eat more fruits and vegetables, than they customarily do, as these food items are presumed healthy. However, the reason given by the participants for insufficient registration in the end of the recording period was that they found the keeping of records for more than 4 days too demanding. Fruits and vegetables are consumed almost every day, and 4-day food record might be sufficiently long time period to represent the usual intake. The correlation coefficients between intake according to the food frequency part of the Questionnaire and the 7-day food record or 4-day food record are comparable to other validation studies on fruit and vegetable consumption

6 Table 6 Cross-classification of participants by quartiles of calculated intake from the 7-day food record and the food frequency part of the Questionnaire, from the 4-day food record and the food frequency part of the Questionnaire and from the 7- and the 4-day food records (n ¼ 36 a ) FFQ-FR (7d) FFQ-FR (4d) FR (7d)-FR (4d) Grossly misclassified (%) Correct plus adjacent quartiles (%) Correctly classified (%) Grossly misclassified (%) Correct plus adjacent quartiles (%) Correctly classified (%) Grossly misclassified (%) Correct plus adjacent quartiles (%) Correctly classified (%) Fruits Vegetables Fruits and vegetables FVJ b a n ¼ 36, except for vegetables as one participant did not fill in the food frequency part for vegetables. b Intake of fruits, vegetables and fruit juice. among adults (Ling et al., 1998; Andersen et al., 2002; Van Assema et al., 2002). The correlation coefficients between the food frequency part of the Questionnaire and the 7-day, as well as 4-day, food record were all within an acceptable range, being lowest for vegetables separately, r ¼ 0.45 and 0.40, respectively. The food frequency part seems therefore less suitable for measuring vegetables alone than fruit or total fruit and vegetable intake, which is similar to earlier results (Andersen et al., 2002). Cross-classification is well suited for seeing whether two methods rank results in a similar or dissimilar manner. The distribution of fruit and vegetable intake was acceptable to be divided into equal quartiles. It was not possible to rank the subject into quartiles according to their juice intake as the distribution was skewed. The number of participants classified in the same quartile according to the food frequency part of the Questionnaire and the 7-day food record was highest for fruits, 94% in same or adjacent quartile, and a bit lower for vegetables, 80%. The 4-day record gave similar results. These values are comparable with those from other studies (Ling et al., 1998). As the 4-day record ranked individuals similar to the 7-day record, a 4-day food record might be sufficient when validating food frequency questionnaires for fruits and vegetables. In the present study overestimation of fruits and vegetables by the food frequency questionnaire did not seem to be a problem. The list of food items in the present study was kept short. Detailed food frequency questionnaires with many food items yield a higher frequency of food consumption than shorter lists; therefore, generalized categories may increase the validity of the questionnaire (Ling et al., 1998; Amanatidis et al., 2001; Andersen et al., 2002). Investigators have stated that overestimation may especially be magnified for fruits and vegetables, which are perceived as healthy and socially acceptable food (Amanatidis et al., 2001). The age of the participants in the present validation study was chosen to be similar to the age of parents to 11-year-old children, this was done since the instrument is meant to be used across Europe for parents to 11-year-old school children. A limitation of this study might be the relatively small sample size and the recruitment of participants through advertisement, which yields self-selected participants. People who are willing to participate in a study on diet in relation to health may be more sensitive to dietary guidelines (Riboli et al., 1997). This might explain a somewhat higher intake of fruit and vegetables in the present study than observed in the last national survey, that is, 230 g/ day (Steingrimsdottir et al., 2003) versus 311 g/day according to the 24-h recall in the present study. The method used in the national survey was one telephone administered 24-h recall on the total diet, not especially designed for fruits and vegetables. The validity of the instrument tested in the present study is considered generalizable. The sample size is most likely sufficient even though the likelihood of nonsignificant difference is possible. Significance of results was tested with nonparametric tests additionally to the 413

7 414 parametric, as the normal distribution had been verified by Kolmogorov Smirnov, which might be regarded as nonsensitive for small samples. Nonparametric and parametric methods showed the same results. A diet much higher in fruits and vegetables than is observed in the Icelandic diet is recommended. The recommendation, that is, 500 g of vegetables, fruits and fruit juice is far from the average intake in Iceland. Median intake is even lower. It is therefore important to increase the intake of fruits and vegetables in the population. The precoded 24-h recall part of the Questionnaire proved to be valid for assessing the mean group intake of fruits and vegetables among adults in a population of low intake, and the food frequency part of the Questionnaire was valid for ranking individuals according to their usual fruit and vegetable intake. The parts then complement each other, and the instrument is therefore valid for assessing fruit and vegetable intake among adults with low intake. Acknowledgements This study has been carried out with financial support from the Commission of the European Communities, specific RTD programme Quality of Life and Management of Living Resources, QLK Promoting and Sustaining Health through Increased Vegetable and Fruit Consumption among European Schoolchildren (Pro Children). It does not necessarily reflect its views and in no way anticipates the Commission s future policy in this area. The Pro Children consortium consists of the following partners: Knut-Inge Klepp (Coordinator), Department of Nutrition, University of Oslo, Oslo, Norway; Carmen Perez Rodrigo, Unidad de Nutricion Comunitaria, Bilbao, Spain; Inga Thorsdottir, Unit for Nutrition Research, Landspitali University Hospital, Reykjavik, Iceland; Pernille Due, Department of Social Medicine, University of Copenhagen, Denmark; Maria Daniel Vaz de Almeida, Faculdade de Ciências da NutriçãoeAlimentaçãodaUniversidadedoPorto, Portugal; Ibrahim Elmadfa, Institute of Nutrition, University of Vienna, Austria; Jóhanna Haraldsdóttir, Research Department of Human Nutrition, Royal Veterinary and Agricultural University, Copenhagen, Denmark; Johannes Brug, Erasmus Medical Center Rotterdam, Department of Public Health, the Netherlands; Michael Sjöström, Unit for Preventive Nutrition, Karolinska Institutet, Stockholm, Sweden; Ilse De Bourdeaudhuij, Department of Movement and Sport Sciences, Ghent University, Belgium. References Agudo A (2005). Measuring intake of fruit and vegetables. Background Paper for the Joint FAO/WHO Workshop on Fruit and Vegetables for Health, 2004, Kobe, Japan. WHO, Electronic Resource. Amanatidis S, Mackerras D, Simpson JM (2001). Comparison of two frequency questionnaires for quantifying fruit and vegetable intake. Public Health Nutr 4, Andersen LF, Bere E, Kolbjornsen N, Klepp KI (2004). Validity and reproducibility of self-reported intake of fruit and vegetable among 6th graders. Eur J Clin Nutr 58, Andersen LF, Johansson L, Solvoll K (2002). Usefulness of a short food frequency questionnaire for screening of low intake of fruit and vegetable and for intake of fat. Eur J Public Health 12, Cade J, Thompson R, Burley V, Warm D (2002). Development, validation and utilisation of food-frequency questionnaire a review. Public Health Nutr 5, De Bourdeaudhuij I, Klepp K-I, Due P, Perez Rodrigo C, de Almeida MDV, Wind M et al. (2005). Reliability and validity of a questionnaire to measure personal, social and environmental correlates of fruit and vegetable intake in year old children in 5 European countries. Public Health Nutr 18, Eriksen K, Haraldsdottir J, Pederson R, Flyger HV (2003). Effect of a fruit and vegetable subscription in Danish schools. Public Health Nutr 6, FAO statistic database (2004): visited at the Hu FB (2003). Plant-based foods and prevention of cardiovascular disease: an overview. Am J Clin Nutr 78 (Suppl 3), S544 S551. Joshipura KJ, Ascherio A, Manson JE, Stampfer MJ, Rimm EB, Speizer FE et al. (1999). Fruit and vegetable intake in relation to risk of ischemic stroke. JAMA 282, Joshipura KJ, Hu FB, Manson JE, Stampfer MJ, Rimm EB, Speizer FE et al. (2001). The effect of fruit and vegetable intake on risk for coronary heart disease. Ann Intern Med 134, Key TJ, Allen NE, Spencer EA, Travis RC (2002). The effect of diet on risk of cancer. Lancet 360, Kim DJ, Holowaty EJ (2003). Brief, validated survey instruments for the measurement of fruit and vegetable intakes in adults: a review. Prev Med 36, Klepp KI, Perez-Rodrigo C, Thorsdottir I, Due P, de Almeida MDV, Elmadfa I et al. (2005). Promoting and sustaining health through increased vegetable and fruit consumption among European schoolchildren: The Pro Children Project. Ann Nutr Metab 49, Ling AM, Horwath C, Parnell W (1998). Validation of a short food frequency questionnaire to assess consumption of cereal foods, fruit and vegetables in Chinese Singaporeans. Eur J Clin Nutr 52, Ness AR, Powles JW (1997). Fruit and vegetables, and cardiovascular disease: a review. Int J Epidemiol 26, Neuhouser ML, Patterson RE, Kristal AR, Eldridge AL, Vizenor NC (2001). A brief dietary assessment instrument for assessing target foods, nutrients and eating patterns. Public Health Nutr 4, Riboli E, Toniolo P, Kaaks R, Shore RE, Casagrande C, Pasternack BS (1997). Reproducibility of a food frequency questionnaire used in the New York University Women s Health Study: effect of selfselection by study subjects. Eur J Clin Nutr 51, Steingrimsdottir L, Thorsgeirsdottir H, Aegisdottir S (1991). The Diet of Icelanders I. Main findings. Icelandic Nutrition Council, Reykjavik. Steingrimsdottir L, Thorsgeirsdottir H, Olafsdottir AS (2003). The Diet of Icelanders Dietary Survey of The Icelandic Nutrition Council 2002, Main findings Research of Icelandic Nutrition Council V, Reykjavik. Steinmetz KA, Potter JD (1996). Vegetables, fruit, and cancer prevention: a review. J Am Diet Assoc 96, Stram DO, Longnecker MP, Shames L, Kolonel LN, Wilkens LR, Pike MC et al. (1995). Cost-efficient design of a diet validation study. Am J Epidemiol 142, The Norbagreen 2002 study (2003). Consumption of vegetables, potatoes, fruit, bread and fish in the Nordic and Baltic countries. TemaNord 2003: 556, Nordic Council of Ministers, Copenhagen.

8 Van Assema P, Brug J, Ronda G, Steenhuis I, Oenema A (2002). A short dutch questionnaire to measure fruit and vegetable intake: relative validity among adults and adolescents. Nutr Health 16, Willett WC (1990). Nutritional Epidemiology. Oxford University Press: New York. Willett WC (2000). Diet and cancer. Oncologist 5, World Cancer Research Fund/ American Institute for Cancer Research (1997). Food nutrition and prevention of cancer: a global perspective. WCFR/AICR: Washington, DC. World Health Organisation (2003). Diet, nutrition and the prevention of chronic diseases Report of a joint WHO/FAO expert consultation WHO, Geneva. 415